The 7-hydroxylation of steroids has potentially widespread biological significance. At least three categories of steroid 7-hydroxylases may be defined. By far the best understood is cholesterol 7-hydroxylase (cholesterol 7-monooxygenase, EC 1.14.13.17), a putative liver-specific (1) microsomal enzyme (2) and member of the cytochrome P450 gene family that catalyzes the rate-limiting step of bile acid biosynthesis(3) . As such, cholesterol 7-hydroxylase plays a critical dual role as the primary enzyme promoting both cholesterol catabolism and bile acid formation. A second category of 7-hydroxylases includes several recently described nonhepatic enzymes that act on side chain-cleaved steroids. A dehydroepiandrosterone/pregnenolone 7-hydroxylase has been reported in brain and other tissues (4, 5) and in adipose stromal cells(6, 7) , and a testosterone 7-hydroxylase in testis, lung, and kidney(8, 9) . The functional significance of these 7-hydroxylated steroids in nonhepatic tissues has not yet been established. The metabolites may regulate the availability of their biologically active precursors or could have biological activity themselves. For example, Morfin and Courchay (5) have suggested that 7-hydroxylated metabolites of pregnenolone and dehydroepiandrosterone may regulate the immune response in mice.

The existence of a third class of steroid 7-hydroxylases that preferentially metabolize hydroxycholesterols is controversial. It is well established that the liver is able to use hydroxycholesterols as well as cholesterol as substrates for bile acid biosynthesis(3, 10, 11) . Recently, it has been suggested that this activity involves a hepatic hydroxycholesterol 7-hydroxylase that is different from cholesterol 7-hydroxylase(12, 13, 14) . Using porcine liver microsomes, Toll et al.(14) were able to separate 25- and 27-hydroxycholesterol 7-hydroxylase from cholesterol 7-hydroxylase activity. Oxysterols such as 25- and 27-hydroxycholesterols (as distinct from cholesterol itself) act via a putative receptor (cf.(15) ) as potent regulators of cholesterol homeostasis, inhibiting 3-hydroxy-3-methylglutaryl-CoA reductase activity and LDL ()receptor levels in vitro(16, 17, 18) . Oxysterols also affect cell growth and viability presumably via their inhibitory effects on cholesterol availability(19, 20) . Given the regulatory importance of hydroxysterols, their further metabolism to 7-hydroxylated products is of considerable interest. Apart from serving as intermediates in bile acid synthesis, 7-hydroxylated hydroxysterols may themselves regulate cholesterol homeostasis. Dueland et al.(21) have made the intriguing proposal that 7-hydroxylase may indirectly induce the LDL receptor gene through inactivation of oxysterol inhibition.

Both 25- and 27-hydroxycholesterols circulate in the blood(22, 23) , and sterol 27-hydroxylase ()mRNA is widely distributed(25) . Both activity and mRNA for 27-hydroxylase have been detected in human and rat ovary(26, 27) . Rat luteal cells can metabolize 25-hydroxycholesterol(28) . 27-Hydroxycholesterol has been shown to inhibit human ovarian cell sterol synthesis(27) .

In this work, we describe a novel ovarian 7-hydroxylase with apparent substrate specificity for 25- and 27-hydroxycholesterols, thus differentiating it from liver cholesterol 7-hydroxylase and from nonhepatic 7-hydroxylases, which catalyze C and C steroids. We further describe the dramatic enhancement of ovarian 7-hydroxylase activity by interleukin-1 (IL-1) and tumor necrosis factor (TNF-), cytokines that have multiple biological effects in the ovary(29, 30, 31, 32, 33, 34, 35, 36, 37) . The physiologic significance of cytokine-induced 7-hydroxylated hydroxycholesterols is unknown at present. However, these studies suggest a possible role for cytokines in the regulation of cholesterol homeostasis in the ovary and perhaps other tissues.

MATERIALS AND METHODS

Reagents and Hormones

2-Hydroxypropyl--cyclodextran (2-HPBCD) was obtained from Pharmatec, Inc. (Alachua, FL) and as a gift from George Reed (American Maize Products Co., Hammond, IN). Recombinant human IL-1 (2 10 units/mg) was generously provided by Drs. Errol B. DeSouza and C. E. Newton (DuPont Merck Pharmaceutical Co.) Recombinantly expressed, naturally occurring human IL-1 receptor antagonist (38) was generously provided by Dr. Daniel E. Tracey (The Upjohn Co.). Recombinant human TNF- was generously provided by Dr. H. M. Shepard (Genentech, South San Francisco, CA). Recombinant human insulin-like growth factor I (IGF-I) was from Bachem California (Torrance, CA).

Authentic cholest-5-ene-3,27-diol (27-hydroxycholesterol) and cholest-4-ene-7,25-diol-3-one (7,25-dihydroxycholestenone) were the generous gifts of Dr. Norman Javitt (New York University, New York). Cholest-5-ene-3,25-diol (25-hydroxycholesterol) and other steroids were obtained from Steraloids, Inc. (Wilton, NH). 25-[26,27-H]Hydroxycholesterol (83 Ci/mmol), [1,2-H]cholesterol (51.7 Ci/mmol), [1,2-H]testosterone (42.5 Ci/mmol), and 25-[16,17-H]hydroxycholesterol (custom preparation) were obtained from DuPont NEN.

Animals and Cell Culture

Solubilization of Steroids

Cellular Radiolabeling Studies

7-Hydroxylase Assay

HPLC

Radioactivity was detected on line by a Flo-One/Beta detector (Packard Instruments), and absorbance at 240 nm was simultaneously monitored by a flow-through spectrophotometer (Lambda-Max, Waters). The column was calibrated with more than 35 steroids(40) , including most of those expected to be found in the ovary. The column tends to separate steroids based on the number of their hydroxyl groups, with less polar steroids eluting first.

Purification of Novel 25-Hydroxycholesterol Metabolites (X1 and X2)

GC/MS

NMR

Liquid Hybridization/RNase Protection Assay

Data Analysis

RESULTS

IL-1-stimulated Metabolism of 25-Hydroxycholesterol to Novel Polar Metabolites in the Rat Ovary

Figure 1: Polar hydroxycholesterol metabolites in granulosa cells: stimulation by IL-1. Granulosa cells (5 10 cells/ml) were initially cultured for 72 h in the absence or presence of IL-1 (50 ng/ml), with and without FSH (2 milliunits/ml). Cells were then washed, and 25-[H]hydroxycholesterol (0.2 µM) plus the same treatment protocols were added for an additional 24 h. Media steroids were extracted and analyzed by HPLC as described under ``Materials and Methods.'' Standards are indicated by arrows as follows: Po, progesterone; Pe, pregnenolone; M, 25-hydroxycholestenone; 20-DHP, 20-dihydroprogesterone. X1 and X2 are polar metabolites eluting at 15.7 and 20.4 min, respectively. Chromatograms from the same representative experiment are shown. The inset in the secondpanel shows the IL-1 dose-dependent formation of X1 and X2 (mean ± S.E., n = four values from two separate experiments).

The IL-1-induced metabolism of 25-hydroxycholesterol to polar metabolites was also apparent in comparable experiments using whole ovarian dispersates (Fig.2). Significant quantities of X1 accumulated in the absence of exogenous IL-1 (Fig.2, toppanel). The formation of X1 and X2 by control cells was reduced (Fig.2, secondpanel) by the concomitant application of a naturally occurring IL-1 receptor antagonist(38) , suggesting that the formation of the polar metabolites in control cells could be related to endogenous production of IL-1. We have observed similar evidence of receptor-mediated, endogenous IL-1 activity for other end points of IL-1 activity in whole ovarian dispersates (e.g. nitrite and prostaglandin production; cf.(33) ). The accumulation of both X1 and X2 was markedly increased above control levels (2.9 ± 0.5-fold, p < 0.05, n = 4) by the addition of IL-1 (Fig.2, thirdpanel). IL-1-induced formation of the X2 metabolite was consistently more prominent in whole ovarian dispersates compared with granulosa cells. The accumulation of X1 and X2 was time-dependent in both whole ovarian dispersates (Fig.2, secondpanel, inset) and granulosa cells (not shown). The addition of receptor antagonist to cells that were exogenously stimulated by IL-1 dramatically inhibited the formation of the polar products (Fig.2, bottompanel), suggesting that the hydroxycholesterol metabolizing action of IL-1 is mediated via its receptor. Data from replicate experiments are summarized in Fig.2(bottompanel, inset), demonstrating that the addition of receptor antagonist promotes a 51% (p < 0.05) and a 66% (p < 0.005) reduction in the formation of X1 and X2 by control and IL-1-stimulated cells, respectively.

Figure 2: IL-1-stimulated polar hydroxycholesterol metabolites in whole ovarian dispersates: inhibition by receptor antagonist. Whole ovarian dispersates (2.5 10 cells/ml) were initially cultured for 72 h in the absence or presence of IL-1 (50 ng/ml), IL-1 receptor antagonist (RA; 5 µg/ml), or trilostane (3 10M). Cells were then washed, and 25-[H]hydroxycholesterol (10 µM) plus the same treatment protocols were added for an additional 24 h. Media steroids were extracted and analyzed by HPLC. Steroid elution times are indicated by arrows as described for Fig.1. Chromatograms from the same representative experiment are shown. The inset in the secondpanel shows the time-dependent formation of X1 and X2 in control (brokenlines) and IL-1-treated (solidlines; mean ± difference, n = two values from one experiment) cells. The inset in the bottompanel shows the accumulation of X1 plus X2 in control, IL-1-treated, and IL-1 + trilostane (IL+T)-treated cultures in the absence (solidbars) and presence (hatchedbars) of receptor antagonist. Results are the mean ± S.E. or difference of two to five separate experiments. Significance by ANOVA analysis, relative to IL-1 in the absence of receptor antagonist, is indicated (*).

The formation of X1 and X2 by IL-1 was also inhibited (47%, p < 0.05) when ovarian cells were cultured in the presence of trilostane (Fig.2, bottompanel, inset), an agent that reportedly inhibits 3-hydroxysteroid dehydrogenase/isomerase activity(47) . Furthermore, a UV peak (characteristic of -3-oxosteroids) was always coincident with formation of the radiolabeled X1 metabolite in both granulosa cells and whole ovarian dispersates. Taken together, these data suggest that the formation of X1 involves, in part, oxidation of the C-3 moiety of 25-hydroxycholesterol by the abundant 3-hydroxysteroid dehydrogenase/isomerase activity present in these ovarian cells. A 3-oxo moiety on X1 was later confirmed as described below (cf.Fig.5and Fig. 7). However, inhibition of X2 formation by trilostane is not explained by trilostane inhibition of 3-hydroxysteroid dehydrogenase. Trilostane inhibition of both X1 and X2 is suggestive of a more generalized inhibition, perhaps of the 7-hydroxylase activity we report herein.

Figure 5: Identification of X1 by GC/MS. X1 was purified from ovarian cultures and analyzed by GC/MS. Representative mass spectra for the methyloxime-trimethylsilyl derivatives of isolated X1 (toppanel) and authentic cholest-4-ene-7,25-diol-3-one (7,25-dihydroxycholestenone) (bottompanel) are shown. The molecular ion (M) is indicated at m/z 589, and removal of methyloxime and trimethylsilyl groups is noted by 31- and 90-mass unit losses, respectively. The m/z 131 ion is composed of the terminal three side chain carbons with the derivatized C hydroxyl group.

Figure 7: Coelution of X1 and authentic 7,25-dihydroxycholestenone on HPLC. Authentic 7,25-dihydroxycholestenone (10 nmol) (toppanel) and the 25-[H]hydroxycholesterol metabolites (bottompanel) from IL-1-stimulated whole ovarian dispersates (prepared as described for Fig.2) were analyzed by HPLC. Elution was simultaneously monitored by absorbance at 240 nm (solidlines) and on-line scintillation counting (brokenline).

Characterization of Polar Hydroxycholesterol Metabolites in the Rat Ovary

Figure 3: Polar hydroxycholesterol metabolites in whole ovarian dispersates: peptide specificity. Whole ovarian dispersates were cultured and analyzed as described for Fig.2, except in the absence or presence of TNF- (10 ng/ml) or IGF-I (50 ng/ml). The doses used were maximally effective in modulating ovarian steroidogenesis (not shown). Results were replicated in an additional experiment. In three other similar experiments (not shown) using granulosa cells rather than whole ovarian dispersates, TNF- stimulated accumulation of X1 and X2. M, 25-hydroxycholestenone.

To determine whether the ovarian metabolism of 25-hydroxycholesterol is sensitive to P450 enzyme inhibitors, granulosa cells were cultured in the absence or presence of IL-1 with and without ketoconazole (5-15 µM), a well-known inhibitor of cholesterol hydroxylases and other P450 enzymes(17, 49, 50, 51) , or with aminoglutethimide (15-30 µM), a potent inhibitor of the P450 enzyme(52, 53) . As shown in Fig.4, IL-1-dependent formation of X1 and X2 was inhibited by ketoconazole in a dose-dependent manner, with nearly complete inhibition at the 15 µM dose. Simultaneous assessment of cell viability by tetrazolium dye reduction (54) indicated that ketoconazole was not toxic to these cells at the doses used (not shown). In contrast to the inhibitory action of ketoconazole, aminoglutethimide had no effect on the formation of X1 and X2. In a parallel experiment (not shown), FSH-dependent P450 activity was inhibited to control levels by aminoglutethimide (15 µM), as expected. These data demonstrate that the ovarian enzyme that catalyzes the formation of X1 and X2 is ketoconazole-sensitive and aminoglutethimide-insensitive.

Figure 4: Formation of polar hydroxycholesterol metabolites: effect of P450 inhibitors. Granulosa cells (5 10 cells/ml) were initially cultured for 72 h in the presence of IL-1 (50 ng/ml), with and without ketoconazole (5-15 µM) or aminoglutethimide (15 and 30 µM). Cells were then washed, and 25-[H]hydroxycholesterol (0.2 µM) plus the same treatment protocols were added for an additional 24 h. Accumulation of X1 (solidlines) and X2 (brokenlines) was determined by HPLC analysis. Data are the mean ± S.E. or difference (n = two to five separate experiments).

Identification of IL-1-stimulated Polar Metabolites by GC/MS and HPLC

Figure 6: Identification of X2 by GC/MS: comparison with reduced X1. X2 was purified from ovarian cultures and analyzed by GC/MS. The comparison spectrum for the trimethylsilyl derivatives of X2 (topsection) and NaBH-reduced X1 (bottomsection, inverted) is shown. The molecular ion (M) is at m/z634, and the base peak is at m/z 544. The representative spectrum of X2 (topsection) is identical to that of authentic cholest-5-ene-3,7,25triol (7,25-dihydroxycholesterol; cf.(14) ). Removal of each trimethylsilyl group is noted by a 90-mass unit loss.

The spectrum for the methyloxime-trimethylsilyl derivative of X1 (Fig.5, toppanel) was characterized by a molecular ion at m/z 589 and ions formed by loss of the oxime (m/z 558 (M - 31)) and trimethylsilyl (m/z 468 (M - 31 - 90); m/z 378 (M - 31 - 90 - 90)) groups, among others. These data indicate a cholestenediolone structure with an underivatized molecular mass of 416 Da.

The spectrum for the trimethylsilyl derivative of X2 (Fig.6, topsection of comparison spectrum) demonstrated it to be a cholestenetriol with a derivatized molecular mass of 634 Da to include three trimethylsilyl groups. The underivatized molecular mass would be 418 Da. The most prominent ion in the spectrum was at m/z 544 (M - 90), although M - 90 - 90 and M - 90 - 90 - 90 ions were also present. The high abundance of the M - 90 ion immediately gave a strong indication of the presence of a 7-hydroxy moiety since virtually all steroid classes (androstanes, pregnanes, cholestanes) give base peaks at M - 90 for trimethylsilyl derivatives of 7-hydroxy compounds. Furthermore, comparison with standards available at the time of analysis indicated that the additional hydroxyl group on X2 was not at C-23, C-24, or C-26.

Sodium borohydride reduction of X1 gave two compounds following GC with similar spectra to X2, but slightly shorter and longer retention times, respectively. These almost certainly represented reduction of the carbonyl group to - and -hydrogens. The comparison spectra for reduced X1 (Fig.6, bottomsection, inverted) and X2 (topsection) show that the only difference is the presence of an ion at m/z 196 in the spectrum for reduced X1. This ion is very distinctive for trimethylsilyl derivatives of steroids with 3,7-dihydroxy-4-ene structures(56) . The presence of a 4-ene group in reduced X1 strongly indicated that the moiety prior to reduction was a 3-carbonyl. The lack of the m/z 196 ion in X2 with an otherwise identical spectrum to reduced X1 indicated that X2 had a 5-ene group.

Subsequent analysis by NMR (see below) definitively identified X1 as cholest-4-ene-7,25-diol-3-one (7,25-dihydroxycholestenone). As shown, isolated X1 has an identical spectrum to that of authentic 7,25-dihydroxycholestenone (Fig.5, bottompanel). Furthermore, the X1 metabolite of 25-[H]hydroxycholesterol from IL-1-stimulated whole ovarian dispersates (Fig.7, bottompanel, brokenline) coeluted with authentic 7,25-dihydroxycholestenone (toppanel) in our HPLC system. Note that only X1 and 7,25-dihydroxycholestenone show the UV absorbance (solidlines) that is typical of -3-oxosteroids. NMR analysis (see below) identified both X1 and X2 as specifically 7 (rather than 7 or 7-keto) metabolites. The X2 metabolite was identified by NMR as cholest-5-ene-3,7,25-triol (7,25-dihydroxycholesterol). Isolated X2 has a spectrum (Fig.6, topsection) identical to that reported for authentic 7,25-dihydroxycholesterol (cf.(14) ). Taken together, these data suggest that X2 is a 7-hydroxy metabolite of 25-hydroxycholesterol that is additionally oxidized at C-3 to form X1.

Identification of IL-1-stimulated Novel Ovarian Metabolites as 7-Hydroxylated Hydroxycholesterols by NMR

The X2 metabolite was identified by NMR as cholest-5-ene-3,7,25-triol. At the time of analysis, a suitable reference compound (i.e. one containing the 5-ene-3,7-dihydroxy structure) was not available. Assignment was further complicated by the presence of an impurity (60%), at a similar concentration to the steroid, which was probably some type of carbohydrate. The characteristic signals for this compound are listed in Table2. As with X1, it was possible to assign chemical shifts to all of the ring protons.

The identification of X1 and X2 as specifically 7-hydroxy is confirmed by the characteristic multiplet pattern and chemical shift (cf.(44) ) of the 7-proton in the NMR spectrum. A 7-proton in a 7-hydroxy compound has a different pattern and shift. A 7-keto compound would show no 7-proton signal. That the relevant signal in each spectrum is a 7-proton is confirmed by the cross-peaks present in the two-dimensional COSY spectrum.

Characterization of IL-1-stimulated Ovarian 7-Hydroxylase Activity

Figure 8: IL-1-stimulated 7-hydroxylase activity in whole ovarian dispersates. Whole ovarian dispersates (5 10 cells/ml) were cultured for 72 h in the absence (solidbar) or presence (hatchedbars) of IL-1 (50 ng/ml). Cells were then sonicated and assayed for 7-hydroxylase activity using 25-[H]hydroxycholesterol (0.2 µM) as substrate. Heat-treated sonicates (100 °C, 10 min) or NADH (0.5 mM)-substituted reactions (rather than NADPH) monitored basal assay activity as indicated (mean ± S.E., n = three to five separate experiments). Significance by ANOVA analysis, relative to maximal activity (IL-1, standard conditions), is indicated (*). The inset is a representative experiment showing a double reciprocal plot of activity (V; picomoles/hour/10 cells) for IL-1-treated cells in the presence of different concentrations of 25-hydroxycholesterol (S; micromolar).

Ovarian 7-Hydroxylase Is Specific for Hydroxycholesterols

Figure 9: IL-1-stimulated 7-hydroxylase activity in whole ovarian dispersates: substrate specificity. Whole ovarian dispersates were cultured in the presence of IL-1 and assayed for 7-hydroxylase activity as described for Fig.8. Activity was measured in the absence (solidbar) or presence (hatchedbars) of a 10-fold excess of unlabeled steroids (2 µM), which were potential competitors of the activity determined using 25-[H]hydroxycholesterol as substrate. Data are normalized relative to activity in the absence of excess steroid (solidbar, 100%), and significance by ANOVA analysis relative to this value is indicated (*). 25-OH, 25-hydroxycholesterol; 27-OH, 27-hydroxycholesterol; Chol, cholesterol; Pe, pregnenolone; Po, progesterone; Testo, testosterone; DHEA, dehydroepiandrosterone. Data are the mean ± S.E. or difference from two to three separate experiments, each in duplicate.

To more directly examine the specificity of ovarian 7-hydroxylase, whole ovarian dispersates were initially cultured for 72 h in the absence or presence of IL-1, after which cells were labeled with [H]cholesterol or [H]testosterone (Fig.10). No polar metabolites of cholesterol (indicative of 7-hydroxylase activity) were observed in either the absence (brokenlines) or presence (solidlines) of IL-1, the chromatograms being qualitatively similar. An increased accumulation of an unidentified, less polar metabolite was consistently observed in IL-1-treated cells. Importantly, in parallel experiments using FSH-pretreated cultures, we observed substantial metabolism of [H]cholesterol to P450 products (not shown), demonstrating that solubility/cellular availability of the added cholesterol was not a problem.

Figure 10: IL-1 does not promote ovarian 7-hydroxylation of cholesterol or testosterone. Whole ovarian dispersates (2.5 10 cells/ml) were initially cultured for 72 h in the absence (brokenlines) or presence (solid lines) of IL-1 (50 ng/ml), after which cells were washed, and either [H]cholesterol (3 µM; upperpanel) or [H]testosterone (3 µM; lowerpanel) was added for an additional 24 h. Media steroids were extracted and analyzed by HPLC. Elution times of standards are indicated by arrows as follows: A1, androstanedione; A2, androstenedione; A3, androsterone and dihydrotestosterone; DHEA, epiandrosterone; Diols, androstenediol, 3-androstanediol, and 3-androstanediol; 7-Testo, 7-hydroxytestosterone. Representative chromatograms are shown.

As shown in Fig.10(bottompanel), a variety of [H]testosterone metabolites were observed in both control (brokenline) and IL-1-treated (solidline) cells. However, no metabolism to 7-hydroxytestosterone was apparent. These data directly confirm the conclusion of the enzyme competition assays (Fig.9), that ovarian 7-hydroxylase is specific for hydroxycholesterols.

To begin to delineate the substrate specificity of ovarian 7-hydroxylase at the molecular level, total RNAs from immature female rat liver and from cultured whole ovarian dispersates were probed with rat liver cholesterol 7-hydroxylase [P]UTP-labeled riboprobe using a highly sensitive and specific liquid hybridization/RNase protection assay (Fig.11). A strong signal (protected fragment) of the appropriate size (346 nucleotides) was apparent for the rat liver positive control. In contrast, no cholesterol 7-hydroxylase mRNA transcripts were apparent for either control or IL-1-treated ovarian cells in spite of an obvious protected fragment for the normalizing probe (RPL19, 153 nucleotides) in both these lanes. (The band seen in the control ovary lane is undigested RPL19.) Given the sensitivity inherent in the liquid hybridization/RNase protection assay and since we were able to detect mRNA transcripts for ovarian 5-reductase (not shown; an enzyme of comparable abundance in this system), it seems likely that immature rat ovaries lack liver cholesterol 7-hydroxylase transcripts.

Figure 11: Analysis of cholesterol 7-hydroxylase transcripts in rat liver and ovary. Whole ovarian dispersates (1.5 10 cells/3 ml) were cultured for 72 h in the absence or presence of IL-1 (50 ng/ml). Thereafter, total RNA was prepared from cultured cells (RNA from 1.5 10 cells, each lane) and from immature female rat liver (20 ng/lane) and subjected to liquid hybridization/RNase protection assay. Both 7-hydroxylase and RPL19 (to monitor equality of loading) transcripts were simultaneously probed. Also shown are nucleotide (nt) size markers (Ambion Inc., Austin, TX) and the undigested, full-length probes for 7-hydroxylase and RPL19. Results were replicated in an additional experiment.

DISCUSSION

Herein, we describe a novel, cytokine-regulated, nonhepatic hydroxycholesterol 7-hydroxylase. Activity was initially observed in cultures of rat granulosa cells or whole ovarian dispersates as the IL-1-dependent, receptor-mediated accumulation of two unknown polar metabolites (X1 and X2) of 25-[H]hydroxycholesterol. X1 and X2 were subsequently identified by GC/MS, NMR, and HPLC as cholest-4-ene-7,25-diol-3-one (7,25-dihydroxycholestenone) and cholest-5-ene-3,7,25-triol (7,25-dihydroxycholesterol). Although autoxidation of cholesterol at C-7 (primarily to 7-oxo or 7-hydroxyl moieties; cf.(58) ) is well known, it is unlikely that the hydroxycholesterol 7-hydroxylation described herein is a nonenzymatic event. IL-1-stimulated ovarian 7-hydroxylase activity demonstrated all the hallmarks of a typical enzyme reaction. It displayed coenzyme specificity (NADH is not effective), Michaelis-Menten kinetics, and heat instability. Product accumulation was time- and IL-1 dose-dependent and did not occur in the absence of living cells. Furthermore, ovarian 7-hydroxylase activity was subject to dose-dependent inhibition by ketoconazole, a well-known inhibitor of P450 enzymes(17, 50, 51) , at a dose range similar to that reported for inhibition of rat hepatocyte cholesterol 7-hydroxylase(49) . In contrast, ovarian 7-hydroxylase activity was not inhibited by aminoglutethimide, a potent inhibitor of the P450 enzyme(52, 53) .

The ovarian hydroxycholesterol 7-hydroxylase enzyme differs from hepatic cholesterol 7-hydroxylase (1, 2, 3) since cholesterol does not compete in assays using 25-hydroxycholesterol as substrate and is not metabolized to any polar products and since no mRNA for hepatic cholesterol 7-hydroxylase can be detected in stimulated or unstimulated ovaries. The ovarian hydroxycholesterol 7-hydroxylase enzyme also differs from nonhepatic enzymes that act on side chain-cleaved steroids(4, 5, 6, 7, 8, 9) , based on indirect and direct substrate specificity studies. However, the identity of the physiologically pertinent substrate for ovarian 7-hydroxylase is unknown at this time. 27- and 25-hydroxycholesterols are good candidates since activity and mRNA for 27-hydroxylase have been detected in human and rat ovary(25, 27) , and rat luteal cells can metabolize 25-hydroxycholesterol to reproductive steroids(28) . Furthermore, we have detected 27-hydroxylase mRNA in cultured granulosa cells and whole ovarian dispersates from both control and IL-1-treated cells (not shown). However, no information exists regarding the presence or regulation of such hydroxysteroids in the ovary. It is unclear whether the ovarian 7-hydroxylase preferentially catalyzes a hydroxycholesterol or a hydroxycholestenone as substrate. Certainly, oxidation of hydroxycholesterols at C-3 by 3-hydroxysteroid dehydrogenase/isomerase is prominent in the ovary, as has been shown in the adrenal gland (59) and liver(60) . Thus, any of a variety of 7-hydroxylated hydroxycholesterols or hydroxycholestenones may be of biological significance in the ovary. The challenge in identifying a physiologic role for ovarian 7-hydroxylase will be to determine what are its substrates in vivo.

We have previously reported rat ovarian IL-1 gene expression (61) and a wide range of receptor-mediated (62) biological activities for IL-1 in the ovary(29, 30, 31, 32, 33) . TNF- modulates rat ovarian steroidogenesis(34, 35, 36, 37) . Both these cytokines can also induce hydroxysterol 7-hydroxylase activity in the ovary. This effect is peptide-specific, however, since other ovarian modulatory factors such as FSH and IGF-I (48) are ineffective in this regard. Although this report extends our understanding of the ovarian actions of cytokines, it is perhaps of more interest as a demonstration of the existence of a previously unreported class of 7-hydroxylases that are regulated by IL-1 and TNF-. Cholesterol 7-hydroxylase is primarily regulated at the level of gene transcription via feedback inhibition by bile acids, but is also subject to regulation by diurnal, hormonal, and dietary factors(1, 3, 57, 63) . Regulatory mechanisms for liver hydroxycholesterol 7-hydroxylase or for other nonhepatic 7-hydroxylases are unknown. Therefore, the potential relevance of cytokines to the regulation of steroid 7-hydroxylases is of interest and remains to be determined.

A physiologic function for 7-hydroxycholesterols or 7-hydroxycholestenones can only be speculated upon at this time. Although hydroxycholesterols are substrates for P450 ( (22) and (28) ; cf.Fig.1), preliminary studies (not shown) indicate that 7-hydroxylated hydroxycholesterols are not side chain-cleaved, suggesting that such steroids are not precursors for reproductive steroids. It is possible that 7-hydroxy metabolites mediate some of the biological actions of cytokines in the ovary or play a role in ovarian cholesterol homeostasis. With respect to the latter, Dueland et al.(21) showed that Chinese hamster ovary cells, expressing hepatic 7-hydroxylase, are resistant to down-regulation of LDL receptor genes by 25-hydroxycholesterol. These authors propose that 7-hydroxylase increases the expression of LDL receptor by metabolic inactivation of hydroxycholesterol inhibitors of this gene. This hypothesis also suggests a mechanism by which liver cells (that express 7-hydroxylase) are resistant to down-regulation of LDL receptor(20) . However, the nature of the link between 7-hydroxylase and cholesterol homeostasis is controversial(14, 64) .

Treatments that stimulate macrophages also increase hepatic cholesterol synthesis and mRNA levels for 3-hydroxy-3-methylglutaryl-CoA reductase(65) . Macrophage products (i.e. cytokines) increase serum cholesterol levels(66) . The Dueland hypothesis(20, 21) proposes that cholesterol biosynthesis is stimulated by 7-hydroxylase-mediated inactivation of oxysterol repressors. Perhaps cytokines enhance cholesterol biosynthesis (cf.(66) ) by just such a mechanism. The data described herein are consistent with these previous observations and, taken together, predict that cytokines could enhance cholesterol accumulation in the ovary via the inactivation of inhibitory hydroxycholesterols by cytokine-induced 7hydroxylation.

Although we may speculate that cytokine-induced 7-hydroxylation blocks the inhibitory action of oxysterols on cholesterol biosynthesis, thereby enhancing cholesterol availability, the significance of the IL-1-stimulated 7-hydroxylation of hydroxysterols in the ovary (or other tissues) remains to be established. In any case, the finding of a nonhepatic C 7-hydroxylase that is regulated by IL-1 (and TNF-) suggests a unique role for cytokines in the metabolism of C steroids.

FOOTNOTES

*

This work was supported by a Special Research Initiative Support award and a Frank C. Bressler Research Fund award from the University of Maryland (to D. W. P.) and by National Institutes of Health Research Grants HD-19998 and HD-30288 (to E. Y. A.), HD-O6274 (to J. F. S.), and DK-34400 (to C. S.). NMR work was supported by the University of London Intercollegiate Research Service 600-MHz NMR Service at Queen Mary and Westfield College (London). A preliminary account of this work has been published (Payne, D. W., Tedeschi, C., Strauss, J. F., and Adashi, E. Y. (1992) in 39th Annual Meeting of the Society for Gynecologic Investigation, p. 230, Abstr. 247, Society for Gynecologic Investigation, Washington, DC). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore by hereby marked ``advertisement'' in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

§

To whom correspondence should be addressed: Div. of Reproductive Endocrinology, University of Maryland School of Medicine, 655 W. Baltimore St., Rm. 11-010, Baltimore, MD 21201. Tel.: 410-706-4050; Fax: 410-706-5747.

¶

Present address: Rebecca Sieff Hospital, 1300 Safed, Israel.

The abbreviations used are: LDL, low density lipoprotein; IL, interleukin; TNF-, tumor necrosis factor ; FSH, follicle-stimulating hormone; 2-HPBCD, 2-hydroxypropyl--cyclodextran; IGF-I, insulin-like growth factor I; HPLC, high performance liquid chromatography; GC/MS, gas chromatography/mass spectrometry; RPL 19, rat ribosomal protein L19; ANOVA, analysis of variance.

The sterol 27-hydroxylase is elsewhere referred to as 26-hydroxylase. In this paper, we designate the enzyme as 27-hydroxylase, which identifies both 27-hydroxylase and 26-hydroxylase for reasons described previously (24). Similarly, herein, 27-hydroxysterol designates both 27-hydroxysterol and 26-hydroxysterol.

ACKNOWLEDGEMENTS

REFERENCES

1. Jelinek, D. F., Andersson, S., Slaughter, C. A., and Russell, D. W. (1990) J. Biol. Chem. 265,8190-8197 [Abstract/Free Full Text]
2. Straka, M. S., Junker, L. H., Zaccarro, L., Zogg, D. L., Dueland, S., Everson, G. T., and Davis, R. A. (1990) J. Biol. Chem. 265,7145-7149 [Abstract/Free Full Text]
3. Vlahcevic, Z. R., Heuman, D. M., and Hylemon, P. B. (1991) Hepatology 13,590-600 [CrossRef][Medline] [Order article via Infotrieve]
4. Akwa, Y., Morfin, R. F., Robel, P., and Baulieu, E. E. (1992) Biochem. J. 288,959-964
5. Morfin, R., and Courchay, G. (1994) J. Steroid Biochem. Mol. Biol. 50,91-100 [CrossRef][Medline] [Order article via Infotrieve]
6. Khalil, M. W., Strutt, B., Vachon, D., and Killinger, D. W. (1993) J. Steroid Biochem. Mol. Biol. 46,585-595 [CrossRef][Medline] [Order article via Infotrieve]
7. Khalil, M. W., Strutt, B., Vachon, D., and Killinger, D. W. (1994) J. Steroid Biochem. Mol. Biol. 48,545-552 [CrossRef][Medline] [Order article via Infotrieve]
8. Inano, H., and Tamaoki, B. (1971) Biochemistry 10,1503-1509 [CrossRef][Medline] [Order article via Infotrieve]
9. Sonderfan, A. J., Arlotto, M. P., and Parkinson, A. (1989) Endocrinology 125,857-866 [Abstract]
10. Swell, L., Schwartz, C. C., Gustafsson, J., Danielsson, H., and Vlahcevic, Z. R. (1981) Biochim. Biophys. Acta 663,163-168 [Medline] [Order article via Infotrieve]